Interaction mitochondrial withisolated mitochondria: Mechanism · 2005. 5. 16. · [Mitochondria] (mg mL1) FIG. 2. Curvesforbindingofpresequences. Conditionswerethe same as in Fig.

Proc. Nail. Acad. Sci. USAVol. 89, pp. 608-612, January 1992Biophysics

Interaction of a synthetic mitochondrial presequence with isolatedyeast mitochondria: Mechanism of binding and kinetics of import

(protein translocation/amphiphilic peptides/membranes/Saccharomyces cerevisiae)

DAVID RoISEDepartment of Chemistry, University of California, San Diego, La Jolla, CA 92093-0506

Communicated by Bruno H. Zimm, October 14, 1991 (received for review August 15, 1991)

ABSTRACT The mechanism of interaction of a prese-quence with isolated yeast mitochondria was examined. Asynthetic peptide corresponding to a matrix-targeting signalwas covalently labeled with a fluorescent probe. Binding of thepresequence to the surface of the mitochondria and transloca-tion of the presequence into the interior of the mitochondriacould then be monitored directly in solution by measuringchanges in the steady-state fluorescence of the attached fluo-rophore. The binding step was rapid and reversible. Quanti-tation of the binding under equilibrium conditions suggestedthat the initial association of the presequence with the surfaceof the mitochondria occurred by partitioning of the prese-quence directly into the lipid bilayer of the outer membrane.Subsequent translocation of the bound presequence into themitochondria was monitored by measuring the rate of disap-pearance of presequences sensitive to digestion by added tryp-sin. The efficiency of translocation was high, and the rate of thetranslocation was dependent on the electrical potential acrossthe inner membrane. At physiological concentrations of pre-sequence, the rate displayed first-order kinetics with respect tothe concentration of bound presequence and had a rate con-stant of 0.19 min-' at 200C. Several kinetic models for thetranslocation of the presequence are presented that are con-sistent with the experimental results.

The import of nuclear-encoded proteins into mitochondriarequires that the proteins bind to the mitochondrial surfaceand subsequently be translocated across the membranes ofthe organelle. Experiments with gene fusions have demon-strated that the targeting sequences typically found at theamino termini of imported precursor proteins are solelyresponsible for the recognition of precursors by mitochon-dria, both in vivo and in vitro (1, 2). These import signals,termed presequences, display no common primary structure(3), but model studies with synthetic presequences haveshown that the sequences are surface-active and have astrong affinity for model membranes (4). These studies sug-gested that presequences may interact directly with the lipidbilayer of the mitochondrial outer membrane.The purpose of the current work was to use isolated, intact

mitochondria in the quantitative analysis of the binding andimport of a presequence and to determine the molecularinteractions that are responsible for these events. Previousstudies have demonstrated that radiolabeled synthetic pre-sequences are capable of being imported into isolated mito-chondria, but have not precisely quantified the interactions(5-8). The current study uses a fluorescently labeled syn-thetic presequence to demonstrate that a presequence caninsert directly into the lipid bilayer of the mitochondrial outermembrane and is imported with kinetics that display first-order dependence on the concentration of the lipid-bound

form of the presequence. The studies provide a quantitativemethod to analyze protein binding to and translocation acrossbiological membranes.

MATERIALS AND METHODSSynthesis and Labeling of Peptides. A peptide having the

sequence MLSLRQSIRFFKPATRTLCSSRYLL was syn-thesized and characterized as described (9). The single cys-teine residue was labeled with 5-iodoacetamidofluorescein(Molecular Probes) in 50 mM Tris'HCI, pH 8.0/0.5 mMEDTA/25% (vol/vol) CH3CN. The fluorescent peptide waspurified as a single, sharp peak on a cation-exchange column(Mono S, Pharmacia) by using a gradient from 0.01 to 1.0 Mammonium acetate (pH 7.0) in 25% CH3CN. Under theseconditions, modified and unmodified peptides were com-pletely resolved. Fractions containing the modified peptidewere lyophilized and dissolved in aqueous 50% ethanol.Concentration of the peptide was determined by amino acidanalysis. The peptide was found to contain a single label forevery molecule, based on the extinction coefficient of fluo-rescein (e490 = 75 cm-l mM-1 at pH > 7.0).

Discontinuous Assays of Peptide Import. Yeast mitochon-dria were purified from strain D273-1OB, using the growthconditions and isolation procedures of Daum et al. (10).Mitochondrial protein concentrations were estimated bymeasuring the A280 of mitochondria dissolved in 1% SDS (6280= 2.1 cm2'mg-1). To assay import of the presequence,duplicate samples containing mitochondria (100 jug) in 1 ml ofstandard buffer (0.6M sorbitol/10mM potassium phosphate,pH 7.4/1 mM ATP/2 mM MgCl2/0.05% fatty acid-free bo-vine serum albumin) were placed on ice. Some ofthe sampleswere treated with valinomycin (1 ,ug/ml) to dissipate themembrane potential. Other samples were treated with 1,10-phenanthroline (1 mM) to inhibit an internal, chelator-sensitive presequence peptidase activity. The fluorescentpeptide was added to a final concentration of 50 nM, and thesamples were placed at 20°C for various times. Following theincubations, the samples were chilled, trypsin (50 ,ug) wasadded to one of the duplicates, and the tubes were spun at14,000 x g for 3 min at 4°C to reisolate the mitochondria. Thesupernatants were removed and the pellets were resuspendedin 1 ml of import buffer lacking albumin and ATP. Thesupernatants and resuspended pellets were treated with Tri-ton X-100 (0.1%) and briefly dispersed in a bath sonifier.Samples that had not contained trypsin were then treatedwith trypsin (50 ,ug/ml) and left at room temperature until thepresequence was completely digested. Fluorescence in eachfraction was read on an SLM/Aminco SPF-500 fluorometer(excitation, 490 nm; emission, 525 nm; 5.0-nm bandpass).

Continuous, Spectroscopic Assays of Peptide Import. Con-ditions were the same as those for the discontinuous assaysexcept that the fluorescence of each sample was monitoredcontinuously. The assays were performed in 1 ml of standardbuffer at 20°C in a quartz cuvette. Mitochondria were addedat time zero. Valinomycin and 1,10-phenanthroline, whenused, were added at 30 sec. Import was initiated by addition

608

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

May

6, 2

021

Proc. Natl. Acad. Sci. USA 89 (1992) 609

of the labeled presequence at 1 min. At the end of the assay,trypsin (50 jug) was added to digest all unimported prese-quences.

RESULTSA synthetic peptide that corresponds to the amino-terminal 25residues ofthe yeast cytochrome oxidase subunit IV precursorwas labeled with iodoacetamidofluorescein. After purificationby cation exchange, the fluorescent presequence was used tomeasure binding of the presequence to isolated yeast mito-chondria and subsequent import of the bound presequenceinto the mitochondria. The discontinuous binding and importassays are similar to those typically used to measure import offull-length, radiolabeled precursor proteins (11). Duplicatesamples ofmitochondria were incubated for various times withthe labeled presequence. After treatment of one of the dupli-cates with trypsin to digest unimported presequences, mito-chondria were reisolated by centrifugation, and the fluores-cence in the pellets was measured after they were dissolved indetergent. In samples not treated with trypsin (Fig. 1, 0), thepresequence showed a high level of binding to mitochondriathat was relatively constant over time. Some decrease inbinding was observed in samples not treated with 1,10-phenanthroline, due to partial digestion of the presequence bya metal-dependent protease (unpublished data; see also ref. 8).In contrast, the amount ofpresequence protected from trypsinin energized mitochondria was initially small but increasedover the course of the assay (Fig. 1 Left, *). Depolarization ofthe membrane with valinomycin significantly reduced theamount of presequence protected over time (Fig. 1 Center, *)and showed that import of the presequence was potential-dependent. Treatment of mitochondria with 1,10-phenanthro-line to block degradation ofthe presequence, however, did notaffect the import (Fig. 1 Right, *).The discontinuous assays of presequence import demon-

strated that the fluorescent presequence was competent forimport and that presequence import was accelerated by theelectrical potential of the mitochondrial inner membrane.However, these assays required that the mitochondria bereisolated and dissolved for each time point in order toseparate and assay bound and free material. This additionalstep limits the accuracy of the experiments in determiningkinetic values, particularly if the translocation is not com-pletely blocked during the reisolation step or if the mitochon-

No Inhibitor +Valinomycin +Phenanthroline0.8.

_0

C 0.60

mc 0.40

.)_0.24

0.00 10 20 0 10 20 0

Time (minutes)10 20

FIG. 1. Discontinuous assays of presequence import. Samplescontained mitochondria (100 ug) suspended in 1 ml of standardbuffer. Incubation mixtures contained the fluorescein-labeled syn-thetic presequence at 50 mM. The mitochondria were either un-treated (Left) treated with valinomycin (1 ,g/ml) (Center), or treatedwith 1,10-phenanthroline (1 nM) (Right). The samples were eithertreated with trypsin (o) or left untreated (o) prior to reisolation of themitochondria. Fraction bound refers to the fraction of the totalrecovered fluorescence found in the resuspended mitochondrialpellet after dissolution with Triton X-100. The results were notcorrected for rapidly reversible binding to the walls of the tubes, sothat values greater than zero at the initial time points in the trypsin-treated samples are due to this background.

dria are damaged during centrifugation and release prese-quences that have been imported. Observation of the fluo-rescence of the presequence directly, however, allowed theconcentrations of the bound, free, and imported forms of thepresequence to be monitored continuously during the incu-bations without the need to reisolate the mitochondria.

Binding of the fluorescent presequence to phospholipidmodel membranes decreases the fluorescence of the prese-quence. This quenching, which is rapid and occurs within themixing time ofthe assay (<10 sec), is caused by the decreasedabsorbance of the fluorophore in the microenvironment of amembrane surface (S. T. Swanson and D.R., unpublisheddata). Quenching of the fluorescein-labeled presequence wasalso observed when the presequence was added to isolatedmitochondria. The quenching was measured at various con-centrations of both presequence and mitochondria to gener-ate binding curves (Fig. 2). The observed curves are consis-tent with a two-state equilibrium between the free, un-quenched presequence and a bound, quenched form of thepresequence. Curves were fit to the data points by using theempirical equation Q = QmaxCM/(Cs2 + CM), where Q is theobserved quenching and CM is the total concentration ofmitochondrial protein (g/liter). The parameter Qmax is aconstant obtained by extrapolating the quenching to infiniteconcentration of mitochondria. It corresponds to the quench-ing experienced by fully bound presequences. The parameterC5M is an empirical parameter that represents the mitochon-drial concentration required for half-maximal quenching. Atthis concentration of mitochondria, halfofthe presequence inthe sample is bound. The value of CSM was found to beindependent of the concentration of the presequence underthe conditions used.The binding experiments showed that the mitochondria

have a high capacity for binding of presequences. At aconcentration of mitochondria equal to C'SM (0.024 g/liter), theconcentration of bound presequence, [PB], is one half thetotal concentration ofpresequence, [PT]. Thus, for [PT] = 100nM, the ratio of bound presequence to total mitochondrialprotein is 2.1 ,mol/g. Since porin, by far the most abundantprotein on the surface of the mitochondrial outer membrane,is present in significantly lower amounts [0.12 ,mol/g infungal mitochondria (12, 13)], there are insufficient outer

0.6

-Co 0.4CQ)

O

r

a

0.2

o.o 0.1 0.2

[Mitochondria] (mg mL1)

FIG. 2. Curves for binding of presequences. Conditions were thesame as in Fig. 1, except that all buffers contained 1 mM 1,10-phenanthroline. Relative quenching is defined as (FT - FbJIFTwhere FT is the fluorescence of the presequence added to 1 ml of asolution containing trypsin (50 ,g) in the absence of mitochondria,and Fobs is the initial fluorescence of the presequence in assayscontaining various concentrations of mitochondria and lacking tryp-sin. Presequence concentration was 50 nM (o) or 100nM (e). A curvedefined by the equation Q = QmaxCM/(Cg? + CM) was fit to thepoints. The best value of Q,,a was 0.57 and the best value of Cg2 was0.024 g/liter for both presequence concentrations.

Biophysics: Roise

Dow

nloa

ded

by g

uest

on

May

6, 2

021

Proc. Natl. Acad. Sci. USA 89 (1992)

membrane proteins to account for the extent of presequencebinding, and the presequence must, at least initially, interactdirectly with the lipid phase ofthe outer membrane. Althoughthe behavior of synthetic amphiphilic sequences in modelsystems suggested that this type of interaction might bepossible (9, 14-16), the results presented here demonstrateclearly that the binding ofa presequence directly to lipids canoccur in a biological membrane. Because of this, the bindingcan be treated as a simple partitioning process (17, 18), wherethe partition coefficient, F (liter/m2), refers to the ratiobetween the concentration of bound presequence (with re-spect to the surface of the mitochondria) and the concentra-tion of free presequence. The partition coefficient is propor-tional to (C00)-1.The kinetics of translocation of the presequence into

mitochondria could also be measured with spectroscopicassays using the fluorescent presequence. Treatment of mi-tochondria with 1,10-phenanthroline inhibits the digestion ofthe imported form of the presequence within the mitochon-dria but does not affect the import process (Fig. 1C). Obser-vation of the fluorescence of the labeled presequence underthese conditions revealed that the quenched fluorescence ofthe bound presequence remained constant during the courseof the assay (Fig. 3), presumably because the imported formof the presequence remained bound to a membrane surfacewithin the mitochondria. Addition of trypsin at various timesduring the assay, however, resulted in the rapid increase ofobserved fluorescence as the external presequences weredigested by the protease and the bound fluorescein wasreleased from the surface of the outer membrane. The size ofthis increase could be used to calculate the concentration ofpresequence remaining untranslocated on the mitochondrialsurface at any time during the reaction, and the kinetics of thetranslocation could thus be followed.A series of import reactions using mitochondria treated

with 1,10-phenanthroline were stopped with excess trypsin atvarious times (Fig. 3). These assays revealed a significantdecrease in the amount of fluorescence released by thetrypsin as the incubations progressed and the presequencewas imported. The concentration of presequence bound tothe surface of the mitochondria at each time was calculatedusing information obtained from the binding experiments.The concentration of presequence bound at the initial time isgiven by [PB]O = [PT]Qo(Qmj-f1, where Q0 is the initial

0.3

cn

U

VL.

0

0.2

0.1

0.0 I0 2 4 6 8 10 12

Time (minutes)

FIG. 3. Time course of protection of the presequence fromtrypsin in mitochondria treated with 1 mM 1,10-phenanthroline.Three separate experiments are superimposed in the figure. Mito-chondria (200 ,ug) were suspended in standard buffer with 1mM ATP.At 1 min, the fluorescent presequence was added to 50 nM. Thereactions were quenched with trypsin (50jSg) at various times duringthe experiments. Breaks in the traces indicate times of the additions.The units of the fluorescence are arbitrary.

quenching. The bound concentration at later times wascalculated from the increase in fluorescence following trypsintreatment (AF). Since the digestion by trypsin is rapidrelative to the import reaction (see Fig. 3) and since therelative molar fluorescence intensities of the free and di-gested forms of the presequence are the same (IF = ID), it canbe shown that the change in fluorescence upon trypsintreatment is given by AF = (IF - IB)[PB]. This equation wasrearranged and used to calculate [PBl. The equation was putin terms of the total fluorescence of the presequence in theabsence of mitochondria (FT = IF[PT]) and the quenching ofthe fully bound presequence (Qm. = (IF - IB)/IF), so that[PB] = [PTIAF/(QmaXFT). Values of[PB] at the initial time andat several later times were determined from fluorescencemeasurements for a series of import experiments performedunder energizing conditions in the presence of 1 mM 1,10-phenanthroline (Fig. 4). The results presented are for twodifferent total concentrations of the presequence at severalconcentrations of mitochondria.

Kinetic Model for Presequence Import. The data from Fig.4 have been fit using a kinetic model illustrated in Fig. 5. Thismodel includes an initial, rapidly established partitioningbetween presequence free in the external solution (PF) andpresequence bound externally to the outer membrane of themitochondria (PB). This binding remains in equilibrium dur-ing the import and is described by the partition coefficient, r.After the presequence binds to the mitochondrial surface, itssubsequent translocation can be considered to be analogousto a unireactant enzymatically catalyzed process, except thatall the steps take place within the confines of the mitochon-drial membranes. Because the overall reaction occurs be-tween two phases, it is important to distinguish betweenconcentrations that are relative to the volume of the bulksolution and those that are relative to the surface area of themitochondrial phase. In the subsequent analysis, the con-centrations ofbound presequences or of intrinsic, membrane-bound proteins that are defined relative to the area of theexternal mitochondrial surface will be noted by a subscript orsuperscript S (for example, [PelS). The concentrations ex-

100A

1 0

i100B

C

0 0

.0

, 10o~~10010

0 2 4 6 8 10

Time (minutes)FIG. 4. Kinetics of the import of the presequence. The time

course of protection from trypsin was monitored as described in thetext. Presequence concentration was 50 nM (o) or 100 nM (0).Mitochondria were at 0.05 g/liter (A), 0.1 g/liter (B), or 0.2 g/liter(C). Lines were generated from the simplified form of the integratedrate equation (Eq. 2) with (k2KEIKs)/Km = 0.19 min-' and CgV =0.024 g/liter. The ordinate is plotted on a logarithmic scale.

610 Biophysics: Roise

Dow

nloa

ded

by g

uest

on

May

6, 2

021

Proc. Natl. Acad. Sci. USA 89 (1992) 611

PFr =

PB's

(PF]

OM PB +E - E- PBk-1

/00, k7N

IM

FIG. 5. Kinetic mechanism for the translocation of a prese-quence. OM, outer membrane; IM, inner membrane; E and E-PB, theunoccupied form of a catalyst of translocation and that catalystoccupied by the presequence, respectively. The partition coefficient,F, is defined in the text. The term P1 represents all imported formsof the presequence. If the peptidase in the matrix is active, theimported presequence is rapidly digested. When the mitochondriaare treated with 1,10-phenanthroline to inhibit the peptidase, most ofthe imported presequence appears to remain membrane-associatedon the extracytoplasmic side of the mitochondrial membranes.

pressed in the two different units are directly related to eachother by a factor KSCM, so that [PB]S = [PB]/KsCM. The valueCM is the concentration of mitochondria (g/liter), and Ks(m2/g) is a proportionality factor that relates the surface areaof the outer membrane of the mitochondria to the amount ofmitochondrial protein.The letter E in the scheme (Fig. 5) describes the unoccu-

pied form of a translocation catalyst, referred to as thetranslocator, and the designation E'PB refers to a complexbetween the translocator and the presequence. In order toformulate an overall rate equation, one must derive anexpression for the concentration of E'PB. From the Briggs-Haldane steady-state assumption (19), Ks = [E]S[PB]S/[E'PB]S = (k-L + k2)/k1. The parameter Ks is a pseudoequi-librium binding constant for the translocator-presequenceinteraction when the concentrations are expressed relative tothe external surface area of the mitochondria. Since [E]s/[E'PBnS = [El/[E-Pn] one can substitute and solve for [E], thebulk molar concentration of the unoccupied translocatpr: [E]= Km[E'PBl/[PBeS. From conservation of mass, [EIT = [E] +[E'PB], where [EIT is the total bulk molar concentration of thetranslocator. Substitution and rearrangement gives [E-PBl =[ElT[PB]s/(KM + [PI]S). The bulk molar concentration of thetranslocator, [EIT, is related to the bulk concentration ofmitochondria by an unknown constant of proportionality, KE(mol of translocator per g of mitochondrial protein), such that[EIT = KECM- Substitution of this into the equation andconversion of [PB]S to the bulk molar concentration of boundpresequence gives [E-PB] = KECM[PB]/(KsCMKS + [PB])-The appropriate rate equation for the measured value of

presequence import is d([PF] + [PB])/dt = -k2[E-PB]. Theterm d([PF] + [PBI)/dt is the instantaneous rate of change ofthe total external molar concentration of the presequencewith respect to time. To fit the experimentally determinedvalues of [PI] as a function of time, an integrated form of thisequation was derived. From the binding experiments, therelationship between [PF] and [PIN is known, and [PFI + [PI]= [PB][(C5/CM) + 1]. Substitution of this expression and theexpression derived above for [E'PBl into the rate equationand rearrangement gives CMKsKsd[PBl/[PB] + d[PBl =-k2KE(CM)2(C50 + CM)-ldt. Intpgration of this expressionover the range [PB]O to [PB] yields

CM4KSln([PB]O/[PB]) + ([PB]O- [PB])

= k2KE(CM)2t/(CO + CM). [1]

The data in Fig. 4 were fit to Eq. 1, with CM = 0.024 g/literand with the terms KsKs and k2KE replaced by singleconstants. No unique convergent fit could be obtained,although the ratio of k2KE to Ks Ks was constant and equalto 0.19 min-'. One explanation for this lack of a unique fit isthat the concentration of the surface-bound form of thepresequence, PB, was below the KS value for the binding stepand the translocator was not significantly occupied. Thiscould normally be tested experimentally by using higher totalpresequence concentrations to increase [PB]S. Higher totalpresequence concentrations, however, cause increased mito-chondrial respiration (see ref. 9) and could result in a de-creased membrane potential or other artifacts.As an alternative to increasing the concentration of the

presequence in the experiments, the same kinetic data couldbe applied to a simplified version of the integrated rateequation that was appropriate for subsaturating PB. In thatcase, [El - [EIT, and [E'PB] could be neglected in theconservation-of-mass relationship. With this change, thederivation was continued as above and yielded the integratedform of the rate equation for subsaturating levels of prese-quence:

K Ksln([PB]O/[PB]) = k2KECMt(C' + CM). [2]

It should be noted that Eq. 2 is the limit of Eq. 1 whenCMKS Ksln([PBlo/[PB]) >» ([PB]0 - [PBl). The data from thepresequence imports gave a good fit to this equation withk2KE(Ks Ks)-' = 0.19 min-1. The lines shown in Fig. 4 weregenerated using this rate constant at appropriate values of[PB]0.

DISCUSSIONThe use of a fluorescently labeled peptide with a sequenceidentical to a naturally occurring presequence that targets aprotein to the mitochondrial matrix has'facilitated the phys-ical analysis of the interactions between such presequencesand intact mitochondria. The three major findings of thiswork are that the presequence can bind directly to the surfaceof the mitochondrial outer membrane, that the import rate ofthe presequence is directly proportional to the concentrationof the presequence on the mitochondrial surface, and that, atleast within the concentration range tested, the transporterdoes not display saturation by the bound presequence.The ability of presequences to bind directly to the surface

of the mitochondrial outer membrane raises questions aboutthe molecular basis for the specific recognition of precursorsby mitochondria in vivo. Because the presequence is trans-lated in the cytoplasm, there should be nothing to prevent itfrom inserting into any membrane that contacts the cyto-plasm. The results presented here, however, suggest that thisinsertion would be rapidly reversible and that precursorsincorrectly bound to other membranes would eventually findtheir way to the mitochondria. Although few studies havedirectly addressed the question of mitochondrial proteinsorting in vivo, the possibility of transient incorrect bindingof precursors to other membranes should not be ignored. Theability of presequences to bind to membranes also suggeststhat the role played by receptors for imported proteins(20-22) may involve interaction with the precursor after itsinitial association with the mitochondria or that these pro-teins may interact with another region of the precursor.The observed kinetics Qf import of the presequence are

consistent with a simple kinetic model for translocation (Fig.5). Within the context of this model, there are at least tworeasonable explanations for the observation that the rate ofimport shows first-order kinetics with respect to the concen-tration ofbound presequence and does not saturate. Both arebased on precedents established for the kinetic description ofunireactant enzyme-catalyzed reactions. In the first expla-

Biophysics: Roise

Dow

nloa

ded

by g

uest

on

May

6, 2

021

Proc. Natl. Acad. Sci. USA 89 (1992)

nation, the Michaelis constant for the bound presequence,K, , is higher than the concentration of presequence that canbe used experimentally without damaging the mitochondria.Thus, the rate of import of the presequence is linearly relatedto the concentration of bound presequence in the range ofconcentrations that are experimentally accessible. This sit-uation is analogous to an enzymatic reaction at relatively lowconcentrations of substrate, where the rate of the reaction isdescribed by an apparent first-order rate constant, kapp =Vmax/Km, and v = Vmax[S]/Km, where [SI is the substrateconcentration. For the mitochondrial case, (k2KE/KS)/Kms isequivalent to Vmax/Km. The value KE/Ks is the total con-centration of the translocator on the mitochondrial surface(mol of translocator per m2 of outer membrane surface) andis a constant for any particular mitochondrial preparation.The expression k2KE/Ks is analogous to kcat[ET for anenzymatic reaction, where kcat is the turnover number and[EIT is the total concentration of enzyme.

In the second explanation, the rate ofthe reaction is limitedby two-dimensional diffusion ofthe bound presequence alongthe mitochondrial surface to the translocation site. Thissituation is analogous to the case of an enzyme reaction thatis limited by the rate of diffusional contact between theenzyme and its substrate and occurs with enzymes that havehighly efficient catalytic steps (23, 24). The effect of this onthe kinetic scheme is that k-1 << k2 (see Fig. 5) and thepseudoequilibrium binding constant becomes Ks = k2/ks. Ifone substitutes this into Eq. 2, the observed first-order rateconstant becomes ksKE/Ks. Again, KE/Ks corresponds tothe concentration of the translocator relative to the mito-chondrial surface.An additional possibility that is not shown in the scheme in

Fig. 5 but that cannot be excluded based on the kineticmeasurements is that the import of the presequence isuncatalyzed; that is, translocation proceeds directly from theform of the presequence bound to the lipids on the surface ofthe mitochondria. In this case, the measured rate constantsimply describes this first-order process. It should be em-phasized that none of the three possibilities described herecan be distinguished on the basis of the kinetic data in Fig. 4,nor can it be concluded that there are no other modelsconsistent with the kinetics. The results do, however, pro-vide a framework for the design of additional experiments todistinguish the various mechanisms.As a final point, the kinetics observed for presequence

translocation may be interesting from the standpoint ofcellular physiology. Enzymes are thought to be under evo-lutionary pressure to increase their values of kcat/Km untiltheir reaction rates become limited by diffusional associationof an enzyme and its substate (23, 24). At the same time, thevalues of Km must remain high enough that the enzymes arenot saturated by substrate under normal physiological con-ditions. The rate-limiting step in the import of proteins intomitochondria in vivo is unknown. Translocation of the pre-sequence, however, must precede the uptake of the rest ofthe precursor protein and is likely to contribute to theobserved kinetics of import. The results presented heredemonstrate that the rate of presequence translocation intoisolated mitochondria does not saturate even at concentra-tions of presequence likely to be higher than those present inthe cytoplasm (25). If the same is true in vivo, and the importsystem is not saturated by precursors, the rate of protein

import into mitochondria would be proportional to the con-centration of precursor proteins in the cytoplasm and couldrespond to changes in those concentrations. Thus, the netflux of precursors into mitochondria could be controlleddirectly by the nucleus at the level of transcription andtranslation rather than remotely by the mitochondria. Addi-tional control of the import flux could also result fromchanges in the mitochondrial membrane potential, if thepotential is coupled to the rate-limiting step of translocation.The models presented here should provide a useful startingpoint to understand the molecular and kinetic mechanisms ofthe translocation of presequences into mitochondria andultimately to describe the overall process of protein import.

I thank Todd Swanson, Andre Brandli, and Jack Kyte for com-ments on the manuscript. I am particularly indebted to Jack Kyte,whose help in suggesting experiments and in interpreting the importkinetics was invaluable. I wish to acknowledge the support andencouragement of Jeff Schatz, in whose lab this work was initiated.The work was partly supported by Grant DCB-9006691 from theNational Science Foundation.

1. Verner, K. & Schatz, G. (1988) Science 241, 1307-1313.2. Hartl, F. U., Pfanner, N., Nicholson, D. W. & Neupert, W.

(1989) Biochim. Biophys. Acta 988, 1-45.3. von Heijne, G. (1986) EMBO J. 5, 1335-1342.4. Roise, D. & Schatz, G. (1988) J. Biol. Chem. 263, 4509-4511.5. Ono, H. & Tuboi, S. (1988) J. Biol. Chem. 263, 3188-3193.6. Glaser, S. M. & Cumsky, M. G. (1990) J. Biol. Chem. 265,

8817-8822.7. Pak, Y. K. & Weiner, H. (1990) J. Biol. Chem. 265, 14298-

14307.8. Furuya, S., Mihara, K., Aimoto, S. & Omura, T. (1991) EMBO

J. 10, 1759-1766.9. Roise, D., Horvath, S. J., Tomich, J. M., Richards, J. H. &

Schatz, G. (1986) EMBO J. 5, 1327-1334.10. Daum, G., Bohni, P. C. & Schatz, G. (1982) J. Biol. Chem. 257,

13028-13033.11. Gasser, S. M., Daum, G. & Schatz, G. (1982) J. Biol. Chem.

257, 13034-13041.12. Freitag, H., Neupert, W. & Benz, R. (1982) Eur. J. Biochem.

123, 629-636.13. Riezman, H., Hay, R., Gasser, S., Daum, G., Schneider, G.,

Witte, C. & Schatz, G. (1983) EMBO J. 2, 1105-1111.14. Tamm, L. K. (1986) Biochemistry 25, 7470-7476.15. Epand, R. M., Hui, W.-H., Argan, C., Gillespie, L. L. &

Shore, G. C. (1986) J. Biol. Chem. 261, 10017-10020.16. Skejanc, H. S., Shore, G. C. & Silvius, J. R. (1987) EMBO J.

6, 3117-3123.17. Schwarz, G., Stankowski, S. & Rizzo, V. (1986) Biochim.

Biophys. Acta 861, 141-151.18. Tamm, L. K. (1991) Biochim. Biophys. Acta 1071, 123-148.19. Segel, I. H. (1975) Enzyme Kinetics: Behavior and Analysis of

Rapid Equilibrium and Steady-State Enzyme Systems (Wiley,New York).

20. Sollner, T., Griffiths, G., Pfaller, R., Pfanner, N. & Neupert,W. (1989) Cell 59, 1061-1070.

21. Kiebler, M., Pfaller, R., Sollner, T., Griffiths, G., Horstmann,H., Pfanner, N. & Neupert, W. (1990) Nature (London) 348,610-616.

22. Pain, D., Murakami, H. & Blobel, G. (1990) Nature (London)347, 444-449.

23. Albery, W. J. & Knowles, J. R. (1976) Biochemistry 15, 5631-5640.

24. Fersht, A. (1977) Enzyme Structure and Mechanism (Freeman,San Francisco).

25. Reid, G. A. & Schatz, G. (1982) J. Biol. Chem. 257, 13056-13061.

612 Biophysics: Roise

Dow

nloa

ded

by g

uest

on

May

6, 2

021